Collision Avoidance Mechanism for Swarms of Drones.

Abstract

This article presents a new approach to collision avoidance in drone swarms, designed for operations in large drone swarms and dynamic environments. The mechanism uses distributed communication, where drones share information about their positions and planned trajectories to predict and avoid collisions. The proposed mechanism enables drones to autonomously cooperate and maintain safe distances in complex scenarios. It is based on the concept of repulsion vectors. The avoidance response is determined by the level of immersion in the protective sphere of obstacles, including other drones. The advantage of the algorithm lies in its simplicity and low computational complexity, allowing it to be used even in small and inexpensive drones. The algorithm was tested in a developed simulation environment, created to handle swarms of over 20 drones and to demonstrate the scalability of the proposed solution. Two scenarios were analyzed: (i) two swarms, each with nine drones, flying on a collision course; (ii) a swarm of 25 drones changing formation. The results showed that the mechanism is effective in avoiding collisions, maintaining safe distances and adapting to changing conditions. The proposed mechanism represents a significant advancement in swarm coordination, offering a robust and scalable solution for real-world applications.